TODO.md

To do

The following things should/could be done sooner or later. Items with question mark are not necessarily a good idea.

• Unscreen phonons of MD driver on supercell (using sparse matrices).
• Set ph.D0 and elph.g0 when generating models with q2r.
• Read Wannier-function centres from _centres.xyz.
• Add long-range argument lr to electron-phonon class elph?
• Save _wsvec.dat together with _hr.dat file (important for small meshes).
• Remember order of atomic species in header of force-constant files.
• Use generic (argument agnostic) function to write all QE input files.
• Use assume_isolated = '2D' in all examples with 2D systems.
• Consider that minima and maxima of bitmaps fall on pixel edges, not centers.
• Use info function from MPI module to raise errors and warnings.
• Disentangle two-dimensional bands with spin degree of freedom.
• Check if cDFPT examples work for for nspin = 2 or noncolin = .true.
• Shift Wannier functions more accurately in STM and STS simulations.
• Move I/O functions and numerical constants to separate modules.
• Free local memory after data has been gathered in global memory.
• Drop Python-2 support in favor of things like (nq, nph, *nk, nel, nel)?
• Replace status arguments etc. by global elphmod(.misc).verbosity.